Waves of thought - Issue #2
How to eliminate AI Slop
5/20/20263 min read
There is a new complaint dominating the internet: the overwhelming abundance of “AI Slop.” Readers are fatigued and professionals have learned to spot machine-written prose with the naked eye. “You can detect if an article was written by AI by looking for specific linguistic patterns, structural tells, and utilizing specialized software tools.” Of course that was written by an AI author. And if you know what to look for it is pretty apparent. That’s not necessarily a bad thing if the content provides value and useful information. Many AI reports are enjoyable and informative. But sometimes it feels like originality is being replaced by algorithms and standardization.
Because thousands of AI-edited manuscripts are hitting their desks, the patterns are becoming very easy to recognize. Now is the time to curb some of that blandness of writing style. Look for these patterns and try to alter your style to be more personality driven. Here are the six main tells that a manuscript was produced by AI.
1. Nonsensical Metaphors
AI models excel at creating mood, but their imagery often collapses under close inspection. Machine prose frequently features metaphors that sound striking and poetic at first glance but make absolutely no logical sense when unpacked. Lose the drama. Example: Describing a quiet moment or a tense room by stating, "Silence was a thunderstorm folded into velvet."
2. Emotional Flatlining
Human prose naturally modulates its emotional register based on the stakes of the scene or topic. AI-generated text tends to remain flat and uniform from section to section. It often maintains a single, heightened dramatic heat without moving through natural human cycles of restraint, escalation, reversal, and release. Does every paragraph feel the same? Example: A casual conversation and a high-stakes conflict scene are written with the exact same robotic tone, vocabulary weight, and dramatic tension.
3. Adjective and Simile Overload
Because AI is trained to be descriptive and comprehensive, it tends to over-embellish. In machine-written articles, nearly every noun is decorated with an adjective, and too many actions are dressed up in complex comparisons. This over-optimization makes the prose feel coded, heavy, and exhausting to read. Example: "The moon hung like a bruised pearl over the trembling, restless, haunted sea as her fractured heart beat like a broken hymn." Do you speak like that? Then don’t write like it.
4. Rigid LLM Rhetorical Patterns
AI engines rely heavily on specific, repetitive dramatic structures to emphasize points. Once you recognize these mechanical habits, they become incredibly easy to spot across blog posts and manuscripts alike. Watch out for predictable syntactical formulas such as:
"It was not [X], it was [Y]..."
"She did not just feel it, she became it..."
"Not [X], not [Y]..."
5. Missing Setting and Sensory Grounding
AI struggles with lived, physical reality. Machine-written text is often heavily reliant on abstract dialogue and vague internal feelings, completely missing concrete scene grounding or physical action. Characters or speakers will debate a topic for paragraphs, but the text fails to establish where they are, how they are moving, or any authentic sensory details of their environment. Be yourself. Set the stage.
6. Over-Reliance on the "Rule of Three"
To make writing sound balanced and rhythmic, AI algorithms constantly sort adjectives and descriptors into neat, rhythmic triplets. When sentences fall into patterned triplets too frequently, the writing begins to sound explicitly assembled rather than naturally discovered.
Examples: "She was furious, frightened, undone" or "The strategy was cold, cruel, calculating." Just scan a few corporate blogs or LinkedIn articles and you’ll see this everywhere.
The Critical Litmus Test for Content Creators
If you use AI in your workflow, you must understand the line between ethical assistance and automated text generation. The core litmus test used by professional editors is simple: Who is writing and editing the prose?
Ethical AI Use: Using an LLM to build outlines, brainstorm concepts, map out content structures, or provide initial developmental feedback.
Unethical/Detectable AI Use: Letting an LLM rewrite your sentences or completely edit your manuscript. When you use a model to rewrite your work because you think "it looks so good," you are actually making your unique voice sound exactly like thousands of other creators using the same tool.
Some suggestions that may help:
Train AI to use more of your personal style when writing. Give it some examples of how you write. The more examples you provide, the better it captures your voice.
Look for the tells above in finished manuscripts. Maybe rewrite or change parts of the document to reflect a more natural tone. Instruct AI to limit the use of these structures.
Verify the text by pasting it into an advanced AI detector like GPTZero, Originality.ai, or Sapling, which analyze the text for uniformity in sentence structure and predictability in word choice.
AI is a great tool for perfecting ideas. It is great for developing outlines, flowcharts, and organizing information (this was written by a human). It can save a lot of development time and help those with non-professional writing skills to correct grammar, punctuation, and sentence structure. AI can be used to evaluate your writing when the prompts are constructed properly. Just don’t rely too heavily on technology. Keeping it human will produce better results in “The Long Run” ( as The Eagles would say).
Dude You Write Like an AI Bot!
Contact
Questions or feedback? Reach out anytime.
© 2026. All rights reserved.
Receive our Newsletter